My colleague Samuel Bowles used to say that there are two types of economist: the Priest and the Engineer. The Priests live in their own little world and spin theories without any reference to the facts. The Engineers live in the real world, collect data, analyze time series, make predictions, give policy advice, and generally ignore all but the most basic economic theory. Certainly, the Engineers never give a thought to what the Priests are doing (they're usually separating hyperplanes, playing with Fredholm operators, or lost in Banach space). Not surprisingly, the microeconomics textbook used in all the best graduate departments around the world has more than a thousand pages stocked with axioms and theorems, but there is not one economic fact in the whole book.
You may ask what this has to do with Vernon Smith's new book, Rationality in Economics. Well, more than a half-century ago, Smith set out to actually test economic theory! He put together a working laboratory, got amazing results having people play games that represent economic exchange, and thus started a movement that has begun to radically change the relationship between fact and theory in economics. Not only has experimental economics (that's what is done in his laboratory) expanded a thousandfold over the years, but leading journals now present models that attempt to account for the observed behavior of human subjects in the laboratory. Vernon Smith received the Nobel prize in economics 2002 for his efforts, but he really deserves half a dozen such awards for his energy, insight, and just play chutzpah.
Vernon Smith changed my life in 1992, when I read an article he wrote in Scientific American surveying his work. To that point I had thought that experimental economics was just a bunch of right-wing know-nothings trying to show Adam Smith's invisible hand really worked, both with humans and pigeons. I was wrong. I can honestly say that everything I assert with confidence about economics comes from either the result of experiments or observing the comparative performance of different real-life economic institutions.
Vernon Smith, in his discussion of the scientific method (Chapter 13) asserts that the idea that we can derive good theory just by carefully looking at the facts is not correct, despite the fact that virtually every great scientist believes he or she does just that. This is a curious position for a man who has spent his life assembling the facts. It is clear that the correct theory does not leap out at us as an ordered kaleidoscope of facts, but the normal everyday application of genius to the facts does quite regularly give rise to better theory. This is why experimental economics is so important.
Intellectually, Vernon Smith is heir to the Scottish Enlightenment, plus Friedrich von Hayek. Hayek plays a front and center role in this book in three or four different ways. The most important really has nothing to do with experimental economics, although Vernon Smith does an excellent job of applying Hayekian insights to experimental methodology. Some historical background is in order. Hayek was a conservative Austrian school aristocrat whose major occupation in life was opposing and discrediting socialism, in an era where many intelligent people believed the triumph of socialism was virtually inevitable. In the mid-1930's, Hayek, Ludwig von Mises, and others entered into a debate with the market socialists Oskar Lange and Abba Lerner as to the feasibility of state socialism. Neoclassical economic theory was accepted by all, and the socialists won a decisive victory. This is not surprising, since neoclassical economic theory accepted the Walrasian general equilibrium model, in which a centralized actor, the auctioneer, orchestrated the whole economy, the only role of markets being to carry out s set of plans architected from the central planner (the auctioneer himself). What different did it make if there is private property and an auctioneer or state ownership of the means of production and a state planning board? According to neoclassical economics, none whatever.
Indeed, as a result of this debate, the great Austrian-American economist Josef Schumpeter wrote a whole book explaining why socialism was inevitable, despite how much he despised it (Capitalism, Socialism, and Democracy, 1942). This defeat led Hayek in quite a different direction. He deeply rethought his commitment to neoclassical economics. In the late 1930's Hayek developed the argument that the economy is an organic entity that cannot be properly modeled analytically, and information is so decentralized and distributed among economic actors that no sort of centralized planning was possible. He was prevented from publishing his ideas until the conclusion of WWII, his article appearing as "The Use of Knowledge in Society", American Economic Review 35,4 (1945):519-530. Read it and weep, socialists.
It would be hard to overstress the importance of Hayek's ideas, and their essential correctness. Now, Hayek himself used his theory as an instrument for denying an important place for the state in regulating the business cycle, providing social insurance, and dealing with market failures in a productive economy. In this, like Milton Friedman after him, he was very incorrect. There is no advanced market economy without an equally advanced, interventionist state. But, his critique of the idea that planning could replace market competition is absolutely correct, and is the idea underlying the subtitle of Vernon Smith's book: "Constructivist and Ecological Forms." Smith offers a dialectical view in which social planners, scientists, and entrepreneurs develop analytical theories (constructs), but only their subjection to the ecology of market competition and the competition of ideas dedicated to the experimental method gives rise to social institutions, products, and ideas that are dynamically robust.
The description and analysis of experimental economics in this book is first rate, as one might expect. The discussion of the spectrum auctions is an especially wonderful exposition of fundamental theory without all the equations. Smith's overall model of human behavior is insightful, and draws creatively on the fact that we are the product of gene-culture coevolution, so our ecological rationality make look weak to the constructivist models of the experimenters, but it is the most advanced form of information processing on this planet by far.
Like most experimental economists, Vernon Smith is unmoved by the contributions of sociological theory. He certainly betrays no interest in the theory of social roles and social actors, or the concept of the internalization of norms. These are, however, among the most important tools in the behavioral scientist's arsenal for understanding exactly how and why humans are not simply selfish sociopaths. A good reading of Emile Durkheim, Talcott Parsons, and a few other major sociological figures would add materially to the insights of the Scottish school.
Morality is a biologically evolved aspect of human genetic constitution, itself a product of the gene-culture coevolutionary process that underlies Vernon Smith's understanding of human nature. People sacrifice material benefits to achieve moral ends, including such character virtues as honesty, trustworthiness, and fairness. I used to doubt the centrality of this biological aspect of our nature, but experimental economics and neuroeconomics have disabused me of any notion that we could ever understand human behavior in terms self-interest, enlightened or otherwise.
Vernon Smith embraces Hayek's model of human psychology in The Sensory Order (1952), which is indeed a brilliant work adumbrating modern connectionism. What is often overlooked is that Hayek's aim was to develop a completely decentralized model of human psychology for the same reasons he preferred a completely decentralized model of the ecological economy. In this, Hayek was really closer to the old associationists of the classical liberal period. although utilizing a relatively sophisticated understand of brain architecture. Like the model of well-known psychologist Donald Hebb, he treated learning as a process of increased communication ease between axons of cells that, based on external stimuli, tended to fire together (Hebb's law is often expressed in simplified form as "Neurons that fire together wire together."
The problem with this view of human psychology is that it is so brain-based that it cannot deal with the emergent properties of human dyadic or multi-adic communication and strategic interaction. There is no difference, for instance, in the brain of a monkey and a human from the Hayekian/Hebbian point of view, except quantitative.
In fact, as I show in my book The Bounds of Reason (Princeton, 2009), it is impossible to understand human strategic interaction without overarching social norms of the type first studied by Emile Durkheim, and now a standard part of the sociological literature. Moreover, modern neuroscience show that there are indeed specialized structures to deal with human interaction and personal morality. This does not contradict Hayek, of course, it is an important addition to the argument in The Sensory Order.